Gravells, P., Neale, J., Grant, E., Nathubhai, Amit, Smith, K.M., James, D.I. and Bryant,   H.E.   (2017)  Radiosensitization  with   an   inhibitor   of   poly(ADP­ribose) glycohydrolase:   A   comparison   with   the   PARP1/2/3   inhibitor   olaparib.   DNA Repair, 61. pp. 25­36. ISSN 1568­7864 

Downloaded from: http://sure.sunderland.ac.uk/id/eprint/9785/

Usage guidelines

Please   refer   to   the  usage guidelines  at  http://sure.sunderland.ac.uk/policies.html  or  alternatively contact sure@sunderland.ac.uk.

Contents lists available at ScienceDirect

DNA Repair

journal homepage: www.elsevier.com/locate/dnarepair

Radiosensitization with an inhibitor of poly(ADP-ribose) glycohydrolase: Acomparison with the PARP1/2/3 inhibitor olaparib

Polly Gravellsa, James Nealea, Emma Granta, Amit Nathubhaib, Kate M. Smithc,Dominic I. Jamesc, Helen E. Bryanta,⁎

a Academic Unit of Molecular Oncology, Sheffield Institute for Nucleic Acids (SInFoNiA), Department of Oncology and Metabolism, University of Sheffield, Beech HillRoad, Sheffield, S10 2RX, United KingdombDrug and Target Discovery, Department of Pharmacy and Pharmacology, University of Bath, Claverton Down, Bath, Somerset, BA2 7AY, United Kingdomc Drug Discovery Unit, Cancer Research UK Manchester Institute, The University of Manchester, Wilmslow Road, Manchester, M20 4BX, United Kingdom

A R T I C L E I N F O

Keywords:PARPPARGTankyraseRadiosensitizationHomologous recombinationNon-homologous end-joining

A B S T R A C T

Upon DNA binding the poly(ADP-ribose) polymerase family of enzymes (PARPs) add multiple ADP-ribosesubunits to themselves and other acceptor proteins. Inhibitors of PARPs have become an exciting and realprospect for monotherapy and as sensitizers to ionising radiation (IR). The action of PARPs are reversed by poly(ADP-ribose) glycohydrolase (PARG). Until recently studies of PARG have been limited by the lack of an in-hibitor. Here, a first in class, specific, and cell permeable PARG inhibitor, PDD00017273, is shown to radio-sensitize. Further, PDD00017273 is compared with the PARP1/2/3 inhibitor olaparib. Both olaparib andPDD00017273 altered the repair of IR-induced DNA damage, resulting in delayed resolution of RAD51 focicompared with control cells. However, only PARG inhibition induced a rapid increase in IR-induced activation ofPRKDC (DNA-PK) and perturbed mitotic progression. This suggests that PARG has additional functions in the cellcompared with inhibition of PARP1/2/3, likely via reversal of tankyrase activity and/or that inhibiting theremoval of poly(ADP-ribose) (PAR) has a different consequence to inhibiting PAR addition. Overall, our data areconsistent with previous genetic findings, reveal new insights into the function of PAR metabolism following IRand demonstrate for the first time the therapeutic potential of PARG inhibitors as radiosensitizing agents.

1. Introduction

The poly(ADP-ribose) polymerase (PARP) family of enzymes arerecruited to, and activated at, sites of DNA damage, where they addpoly(ADP-ribose) (PAR) to themselves and to other DNA repair andchromatin-remodeling factors [1,2]. Once synthesised the PAR polymeris thought to act as a signal to recruit repair factors to the damage. Inthis way PARP proteins are considered to play a key role in co-ordinating the repair of single [3–10] and double strand DNA breaks[11–15], and in the restart of stalled or collapsed DNA replication forks[16–18]. Given this key function in DNA repair, several inhibitors of thePARP proteins are now under development for cancer treatment, to beused either alone [19] or in combination with DNA damaging agentssuch as radiotherapy (reviewed in [20]). PARP1 depletion has beenshown to modestly increase sensitivity to ionising radiation (IR) inmouse models [21,22]. In addition, a variety of PARP inhibitors, re-portedly targeting PARPs 1, 2 and 3 to various degrees, have beendemonstrated to radiosensitize a variety of human tumour cell lines

[23–27] including breast cancer [28–31], and have shown success inseveral preclinical and clinical trials [32–43]. Radiosensitization bythese inhibitors is generally considered to be a replication dependentevent [44,45].

The catalytic action of all PARPs are reversed by the endo- andexoglycosidase activities of poly(ADP-ribose) glycohydrolase (PARG)[46–50], and it is proposed that following recruitment of other repairproteins to sites of damaged DNA, PAR must be removed for DNA repairto be completed [6]. Consistent with a role in DNA repair, PARG de-ficient cells have been reported to display reduced efficiency of doublestrand break (DSB) [51–53] and single strand break (SSB) repair [6],and to have difficulties during situations of replication stress [53–56].These defects in repair/replication suggest that PARG like PARP is apossible target as a single agent in certain genetic backgrounds [53] andfor sensitizing to DNA damaging agents. The reported chemosensitizingeffects are variable [6,52,57–61], but gene depletion or silencing ofPARG using siRNA has consistently resulted in sensitivity to ionisingradiation (IR) in mouse ES cells [62,63] and human cancer cell lines

https://doi.org/10.1016/j.dnarep.2017.11.004Received 7 August 2017; Received in revised form 17 November 2017; Accepted 17 November 2017

⁎ Corresponding author.E-mail address: h.bryant@sheffield.ac.uk (H.E. Bryant).

DNA Repair 61 (2018) 25–36

Available online 22 November 20171568-7864/ © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/BY/4.0/).

T

[51,64], with accumulation of mitotic defects and death occurring bymitotic catastrophe [51,64].

Each of the radiosensitizing studies above was carried out in cellsdeficient in PARG, and while supportive, the investigation of the ther-apeutic potential of PARG has been limited by the lack of a cellpermeable, specific, PARG inhibitor. Recently, we developed a novel,first in class, PARG inhibitor – PDD00017273 [65], which showedsynthetic lethal killing in cells deficient in certain homologous re-combination associated proteins [66]. Here we test the ability of thesame agent to sensitize breast cancer cells to IR. In addition, we com-pare this with the radiosensitizing effects of olaparib. Olaparib has re-ported IC50 values of 5 nM, 1 nM and 4 nM for PARP1, PARP2 andPARP3 respectively [67].

2. Materials and methods

2.1. Cell culture

The MCF-7 and MDA-MB-231 breast epithelial adenocarcinoma celllines were purchased from the American Type Culture Collection(ATCC® HTB-22™ and ATCC® HTB-26™ respectively). Cell lines weremaintained in Dulbecco’s modified Eagle Medium (DMEM, Gibco,ThermoFisher Scientific, MA, USA) supplemented with 1× non-essen-tial amino acids (NEAA, Sigma-Aldrich, MO, USA) and 10% Foetalbovine serum (Gibco) at 37 °C under an atmosphere containing 5% CO2.

2.2. Inhibitors

The PARG inhibitor, PDD00017273, [65] was resuspended in di-methyl sulfoxide (DMSO) at a concentration of 20 μM and stored at−20 °C. A final concentration of 0.3 μM was used. The PARP inhibitor,olaparib, was purchased from Cambridge Biosciences (UK) and pre-pared in DMSO to give a 1000× stock. A final concentration of 1 μMwas used. The dual-site binding tankyrase inhibitor – 8-Methyl-2-(3-oxo-3-(4-((quinolin-8-yl)aminocarbonyl)-phenylamino)propyl)quina-zolin-4-one (compound 14 in reference [68]) was prepared as a 5 mMstock in DMSO. A final concentration of 5 μM was used.

2.3. SiRNA transfection

ON-TARGETplus siRNA was purchased from Dharmacon (GEHealthcare Life Sciences, CO, USA) for two individual PARG(NM_003631) siRNA oligonucleotides, PARP1 (NM_001618) and thenon-targeting siRNA #1 (scramble) control. All siRNAs were re-suspended at 20 μM in 1 × siRNA universal buffer (Dharmacon) andstored at −20 °C. Cells were seeded in 6-well plates and left overnightto attach. The following day, cells were transfected with 20 nM siRNA(final concentration) using Dharmafect 4 reagent (Dharmacon) fol-lowing manufacturers’ instructions. Knockdown was confirmed after48 h by western blotting.

2.4. Clonogenic survival assay

Cells were plated at known densities in 90 mm dishes and left toattach for 4 h. After this time, inhibitors were added to the media at theconcentrations stated above. The next day, cells were exposed to in-creasing doses of IR using an IBL437C Irradiator (Source 51.5TBq,Cs137) and then left for 15 days to form colonies. Colonies were stainedwith 4% methylene blue in 70% methanol and counted. Where siRNAknockdown was used, cells were transfected in 6-well plates and left for48 h before replating at known densities in 90 mm dishes and exposingto increasing doses of IR.

2.5. Western blotting

Cells were lysed in RIPA buffer (50 mM Tris-HCl, 150 mM NaCl, 1%

Triton X-100, 0.1% SDS, 1 mM EDTA, and 1% sodium deoxycholate) inthe presence of 1× protease and phosphatase inhibitor cocktails(Roche, Sigma-Aldrich, MO, USA). An aliquot of 30 μg total protein(measured by BioRad DC protein assay) was run on an SDS-PAGE geland transferred to Hybond ECL membrane (GE Healthcare, CO, USA).This membrane was immunoblotted with antibodies against Poly(ADP-ribose) 10H (1:400, Enzo Life Sciences, NY, USA), PARG (1:500, SantaCruz Biotechnology, TX, USA), PARP1 (1:1000, Santa CruzBiotechnology) and TUBB (β-tubulin; 1:2000, Sigma-Aldrich), eachdiluted in 5% milk and incubated at 4 °C overnight. After the additionof the appropriate HRP-conjugated secondary antibody and furtherwashes, the immunoreactive protein was visualised on Hyperfilm™ ECL(GE Healthcare) using ECL reagents (GE Healthcare) following manu-facturer’s instructions.

2.6. Immunofluorescence

Cells were plated on to coverslips and allowed to settle beforetreating with inhibitors overnight. The next day, cells were irradiated at3 Gy and then either fixed immediately or left to repair at 37 °C for thetime stated in the figures. Cells were fixed in 4% paraformaldehydesolution (Insight Biotechnology Ltd, UK) for 20 min at room tempera-ture and then extensively washed (3 × 5 min in tris-buffered saline(TBS), 1 × 10 min in phosphate-buffered saline (PBS) containing 0.5%Triton X-100 and 3 × 5 min in TBS). Coverslips were placed in 10%goat serum (ThermoFisher Scientific) in TBS for 1 h at room tempera-ture to block followed by a further 3 × 5 min washes in TBS prior toincubation with the primary antibodies anti-γH2AX (ser139) (CellSignaling, MA, USA), RAD51 (Santa Cruz Biotechnology), DNA-PKcspS2056 (Abcam, UK) or poly(ADP-ribose) 10H (Enzo Life Sciences),Pericentrin (Abcam), β-Tubulin (Sigma-Aldrich) or the PAR bindingreagent MABE1016 (Millipore) each diluted (1:500) in TBS containing3% goat serum, for 16 h at 4 °C. The coverslips were subsequentlywashed 4 × 10 min in TBS followed by incubation with the secondaryantibodies, Alexa-fluor 594 goat anti-rabbit IgG (ThermoFisherScientific) or Alexa-fluor 488 goat anti-Mouse IgG (ThermoFisherScientific) diluted in TBS containing 3% goat serum (1:500) for 1 h atroom temperature and finally washed 3 × 5 min TBS. Coverslips werethen mounted onto microscope slides with DAPI containing mountant(Vector Labs, CA, USA).

All images were obtained with a Zeiss LSM 510 inverted confocalmicroscope using planapochromat 63 × /NA 1.4 oil immersion objec-tive and excitation wavelengths 488 nm, 546 nm and 630 nm. Throughfocus maximum projection, images were acquired from optical sections0.5 μM apart and with a section thickness of 1.0 μm. Images wereprocessed using Adobe Photoshop (Abacus Inc.). The frequency of cellscontaining foci was determined by counting at least 100 nuclei on eachslide.

2.7. Identification of mitotic phenotypes

Cells were stained with antibodies against TUBB (β-Tubulin), PCNT(pericentrin) and DAPI and observed by fluorescence microscopy.Mitotic cells were classified into prophase, metaphase, anaphase andtelophase stages based on DAPI staining of the DNA (example images ofclassification are shown in Supplementary Fig. 1). β-tubulin stainedspindle formation was classed as abnormal if it was either monopolar,asymmetric or disorganized. Pericentrin was used to allow identifica-tion of multipolar, monopolar or fragmented centrosomes.

2.8. Micronuclei scoring

Micronuclei were identified by DAPI staining in cells stained forγH2AX in the samples treated with inhibitors alone and 12 h post-IRexposure. Cells with greater than five micronuclei where regarded asnecrotic and therefore not included in the analysis. Micronuclei were

P. Gravells et al. DNA Repair 61 (2018) 25–36

26

then scored as either negative or positive for γH2AX staining andaverage number of micronuclei of either type was calculated from thetotal number of cells counted.

2.9. pH3 staining and cell cycle analysis

Cells were seeded in 90 mm dishes and left to attach for 4 h beforeinhibitors were added to the media. The next day cells were exposed to3 Gy IR and then left for 24 h before fixing in 70% methanol and storedovernight at −20 °C. After washing in PBS cells were resuspended in2 ml PBS supplemented with 0.5% BSA (Sigma-Aldrich) and 0.25%Triton-X100. Following 15 min incubation on ice, cells were re-suspended with Histone H3 pS10 antibody (Abcam, 1:1000) diluted in100 μl of PBS supplemented with 0.5% BSA and 0.25% Triton-X100 andincubated for 2 h. After this time cells were washed with 0.25% Triton-X100 in PBS and then incubated with secondary antibody Alexfluor 488goat anti-mouse IgG (1:100, diluted in 100 μl PBS supplemented with1% BSA) for 30 min protected from light. Following a final wash withPBS, cells were incubated with 5 μl RNaseA (2 mg/ml) and 200 μlpropidium iodide (PI, 50 μg/ml) for 15 min in the dark. Samples wereanalysed by flow cytometry using the FACSCalibur 488 nm laser (BDBiosciences, CA, USA).

2.10. Statistical analysis

Where p values are indicated the Student’s T-test was used foranalysis between two sets of data, in each case two-sided, unpaired testswere carried out.

3. Results

3.1. Both PARP1/2/3 and PARG inhibitors increase sensitivity to ionisingradiation

We previously established that 0.3 μM of the PARG inhibitorPDD00017273 is the optimum dose for inhibition of endogenous PARGactivity, with minimal cell killing in the breast cancer cell line MCF-7[53]. Here, the ability of PDD00017273 to inhibit PARG and lead toaccumulation of PAR was confirmed by western blotting and im-munofluorescent staining (Fig. 1A and B). In contrast, and as expected,incubation with olaparib led to reduced levels of endogenous PAR.Concomitant exposure to olaparib and PDD00017273 also resulted inreduced PAR accumulation, confirming the specificity of the PARG in-hibitor and the reagents.

For therapeutic relevance a relatively low dose of 3 Gy IR waschosen. Ten minutes post-IR increased PAR could be detected usingimmunofluorescence but not by western blotting (Fig. 1A and C).However, in the presence of the PARG inhibitor a large increase in PARwas observed using both techniques, suggesting that PAR synthesis isinduced by IR, but that the turnover is too rapid to always observe incontrol cells. IR induced PAR activity had returned to basal levels10 min post-IR in DMSO treated cells, and remained high in a subset ofcells at 90 mins post-IR when PARG was inhibited (Fig. 1C). Addition ofolaparib or olaparib plus PDD00017273 during IR treatment led toreduced PAR (Fig. 1A). These data demonstrate that inhibition of PARGwith PDD00017273 results in reduced PAR turnover and hence DNAdamage-induced accumulation and persistence of PAR.

Previous genetic studies demonstrated that PARG has potential as aradiosensitizing target, but to date a therapeutic agent has not beenavailable to test this hypothesis. Having demonstrated that PARG in-hibition by PDD00017273 effects PAR turnover following IR, the PARGinhibitor was added to MCF-7 cells 16 h prior to treatment with IR andsurvival determined by clonogenic survival assay (Fig. 2A). PARG in-hibition resulted in approximately 2–3 fold increase in sensitivity to IRcompared with DMSO control (3 Gy – 22% survival vs. 55% in control;p < 0.0001). Consistent with this, depletion of PARG also caused

reduced survival in response to IR (Fig. 2B). The effect induced wassimilar to that seen with olaparib or following co-treatment with bothinhibitors (3 Gy – 15% and 25% survival respectively; p < 0.0001compared with DMSO control). Likewise comparing PARP1 depletedcells with PARG depleted cells, there was no consistent difference in thedegree of radiosensitization induced (Fig. 2B). Western blotting con-firmed siRNA-mediated depletion of PARP1 and PARG (Fig. 2C). A si-milar effect was seen in the triple negative breast cancer (TNBC) cellline MD-MBA-231 (Supplementary Fig. 2). Our data support the ideathat inhibition of PARPs can sensitize to IR in breast cancer, includingboth ER+ and TNBC [28–31]. In addition, we show for the first timethat a specific PARG inhibitor, PDD00017273, can sensitize to IR to asimilar degree in these backgrounds.

3.2. PARP1/2/3 inhibition slows and PARG inhibition speeds up repair ofIR induced DNA double strand breaks

Ionising radiation induces breaks in DNA [69]. Given the proposedfunction of PARG in promoting efficient DNA repair, we examined in-duction and repair of DNA damage. To do this we used phosphorylationof histone H2AX on Ser139 (γH2AX) as a marker of DNA damage, andfollowed the kinetics of repair with time post-IR in the presence orabsence of PDD00017273 or olaparib (Fig. 3). Consistent with ourpublished data [53,66,70] inhibition of PARP1/2/3 or PARG resulted ina significantly increased basal level of γH2AX foci staining (p < 0.05for PARP and p < 0.001 for PARG inhibitors compared with control).As expected, treatment with 3 Gy IR induced a rapid increase in thepercentage of cells displaying greater than 10 γH2AX foci/cell(p < 0.001 at 10 min for each condition compared with correspondingmock irradiated sample). Considering the increased backgroundstaining, the presence of neither inhibitor during irradiation affectedthe initial degree of induction of γH2AX foci. However, the resolution ofγH2AX foci to basal levels during recovery from IR occurred earlier andfaster when PARG was inhibited and was delayed when PARP1/2/3was inhibited. In control cells maximum levels of IR-induced DNA da-mage were reached at 10–30 min, resolution began after 30 min butwas not significant until 6 h post-IR, and basal levels were restoredwithin 24 h post-IR. In the presence of the PARG inhibitor, peak levelswere reached at 10 min post-IR, resolution began between 10 and30 min, was significantly different to control at 1 h post-IR, and basallevels were reached around 12 h after IR. In contrast, olaparib treatedcells showed peak levels of DNA damage at 10 min post-IR, which afteran initial small reduction remained relatively high for up to 6 h withsignificant resolution then occurring between 6 and 24 h after irradia-tion.

Together these data suggest that DNA double strand break repair isaltered by olaparib or PDD00017273 however each agent may notfunction via the same mechanism.

3.3. PARG but not PARP1/2/3 inhibition promotes rapid activation of IR-induced DNA-PKcs, while both PARP and PARG delay resolution of IR-induced RAD51 foci

Radiation induces both single and double strand breaks in DNAhowever the kinetics of repair seen above suggest that it is changes tothe repair of DNA DSBs that accounts for much of the difference be-tween control and PDD00017273/olaparib treated cells. IR-inducedDNA DSBs are repaired by two separate but complementary pathways –non-homologous end-joining (NHEJ) and homologous recombinationrepair (HRR). It is generally considered that repair by NHEJ is rapid,error prone and can occur at all stages of the cell cycle whereas HRR isslower, error free and is restricted to S- and G2-phases of the cell cycle[71]. We examined the relative contribution of each of these pathwaysto the repair of IR-induced DNA damage in the presence/absence ofolaparib or PDD00017273, using activated DNA-dependent protein ki-nase, catalytic subunit (DNA-PKcs, pS2051) foci as a marker of NHEJ

P. Gravells et al. DNA Repair 61 (2018) 25–36

27

and RAD51 recombinase (RAD51) foci as a marker of HRR (Fig. 4). Incontrol cells, DNA-PKcs foci rapidly increased in response to IR, peakedat 30 min and quickly resolved such that only 20% of the foci remainedat 3 h post-IR. The remaining foci were then slowly resolved to basal

levels by 24 h post-IR (Fig. 4A and B). In the same control cells, RAD51foci levels increased slowly to peak at 3 h then decreased to basal levelsby 12 h post-irradiation (Fig. 4C and D). In the presence of olaparib orPDD00017273, the appearance of IR-induced RAD51 foci was delayed

Fig. 1. Inhibition of PARG leads to accumulation andpersistence of poly(ADP-ribose) both alone and followingionising radiation. (A) Protein expression of poly(ADP-ri-bose (PAR), PARG and PARP, following inhibition ofPARG with 0.3 μM PDD00017273, PARP with 1 μM ola-parib, or both, either alone or post 3 Gy ionising radiation(IR). (B&C) Immunodetection of PAR in cells treated withinhibitor and left for various times post-IR as indicated. Inall cases MCF-7 cells were incubated for 16 h with in-hibitor before exposure to IR.

P. Gravells et al. DNA Repair 61 (2018) 25–36

28

compared with control cells but peaked at a similar 3 h post-IR. In in-hibited cells, IR-induced RAD51 foci then persisted at peak levels until6 h (no significant decrease in foci between 3 and 6 h in inhibited cellscompared with a 25% decrease in control cells; p < 0.01), after whichthey were resolved and basal levels were restored by 12 h post-IR(Fig. 4C). In contrast to RAD51 foci, DNA-PKcs foci responded differ-ently depending on whether PARP1/2/3 or PARG was inhibited. In thepresence of olaparib, DNA-PKcs foci increased and peaked at similarlevels, and with similar kinetics, to control cells (approximately 30%cells with greater than 10 foci/cell by 10 min post-IR). However, thesefoci then persisted at peak levels until 3–6 h post-IR (30% and 24% inPARP1/2/3 inhibited cells compared with 11% and 6% in control at 3and 6 h respectively; p < 0.05 and p < 0.001 at each time point)with a significant percentage of cells containing high levels of foci even24 h post-IR (Fig. 4A). On the other hand, in PARG inhibited cells therewas a rapid induction of DNA-PKcs foci to almost double the peak levelseen in control cells (49% compared with 29% respectively; p < 0.01).This increase persisted until 1 h post-IR, after 3 h approximately halfhad been resolved and then similar to controls, basal levels were re-stored by 12–24 h post-IR (Fig. 4A). Together these data suggest thatalthough olaparib and PDD00017273 sensitize to IR, there may be adifference in the way cells respond to IR-induced DNA double strandbreaks, with increased NHEJ activity being prompted in the presence ofPARG inhibitor, but not following PARP1/2/3 inhibition. It is likelythat it is the differential kinetics of activation of each pathway by

olaparib or PDD00017273 following IR that contributes to the differ-ence in resolution of γH2AX foci seen above. Interestingly, in the ab-sence of IR, endogenous levels of RAD51 but not DNA-PKcs foci wereincreased by both inhibitors, suggesting that this PARG inhibitor in-duced NHEJ activity is specific to IR-induced DNA damage.

3.4. PARP and PARG inhibitors specifically increase IR-induced γH2AXpositive micronuclei formation

The formation of micronuclei (MN) can occur due to errors inchromosome segregation during anaphase. IR induces γH2AX positive(+ve) micronuclei in cancer cells, perhaps as a result of unrepairedDSBs and broken chromosome ends being incorporated into IR-inducedMN [72], or reflecting altered chromatin structures caused after ille-gitimate DNA repair [73]. We examined MN formation following IR inthe presence or absence of olaparib or PDD00017273 (Fig. 5 and Sup-plementary Fig. 3). As expected radiation induced total MN formationsignificantly compared with untreated controls, with an average of0.3 MN/cell compared with 0.1 MN/cell in DMSO treated samples(p < 0.001 comparing IR alone to no IR). This increase was greater incells where PARP1/2/3 or PARG were inhibited with IR-induced MNlevels of 0.5 and 0.6 MN/cell for PDD00017273 and olaparib respec-tively (p < 0.05 for each compared with IR alone). When γH2AX po-sitive (+ve) and negative (−ve) MN were analysed the majority of IR-induced MN were −ve. Interestingly though it was the γH2AX +ve

Fig. 2. Inhibition or depletion of PARP or PARG in-creases sensitivity to ionising radiation. (A) Survivalfraction of MCF-7 cells untreated (DMSO), treatedwith PARG inhibitor (PDD00017273), PARP in-hibitor (olaparib), or both for 16 h prior to andduring recovery from ionising radiation (IR). (B)Survival fraction following IR of siRNA transfectedcells as indicated. Survival was measured by clono-genic survival assay. Mean and standard deviation ofthree independent repeats are shown. Statisticalsignificance calculated by Student’s T-test, cf. toDMSO or scrambled siRNA control, where *** re-presents p < 0.001. (C) Protein expression of siRNAtransfected cells 48 h post-transfection.

P. Gravells et al. DNA Repair 61 (2018) 25–36

29

rather than −ve MN that were increased by PARP1/2/3 or PARG in-hibition. This supports the idea that it is altered DNA repair that isresponsible for radiosensitization by olaparib and PDD00017273.

3.5. PARG inhibition has a greater effect on IR-induced metaphaseaberrations than inhibition of PARP1/2/3

Previous reports of the effect of PARP1/2/3 inhibition on IR-in-duced cell cycle distribution are limited however radiosensitization isgenerally thought to be replication dependent [24,27]. In additionthere are conflicting reports regarding the effects of PARG depletion onIR-induced cell cycle checkpoints. For example, enhanced G2/Mcheckpoint arrest is seen in PARG depleted HeLa cells [51], while ab-rogation of this arrest is reported in lung cancer cells [64]. Here, the cellcycle profile of MCF-7 cells following IR in the presence or absence ofolaparib or PDD00017273 was examined by flow cytometry. By co-staining for phosphorylation of H3 Serine 10 (pH3) the percentage ofcells in mitosis could also be observed (Fig. 6). At the relatively lowdoses of radiation used here, in MCF-7 cells the predominant activated

cell cycle checkpoint was the G1/S checkpoint [74]. Addition of thePARG inhibitor PDD00017273 enhanced the G1/S checkpoint arrestwhile reducing the number of cells in S, G2 and M phases of the cellcycle (p < 0.05, compared with equivalent phase in IR-treated non-inhibited cells). Olaparib had a similar effect but with a more pre-dominant effect on S-phase (p < 0.001, compared with IR-treated non-inhibited cells). Neither PDD00017273 nor olaparib altered the cellcycle profile in the absence of IR.

Following IR both PARP and PARG inhibitors reduced the percen-tage of cells in mitosis (Fig. 6B, p < 0.05, compared with IR-treatednon-inhibited cells). The fidelity of mitosis was tested by im-munodetection of microtubules and centrosomes. Following IR in thepresence of PARG inhibition, an increase in the percentage of mitoticcells with aberrant spindle formation was observed (Fig. 7A) includingmonopolar, asymmetric and disorgansied spindle formations (Fig. 7B).This was not seen following inhibition of PARP1/2/3 with olaparib.Aberrant mitosis was accompanied by an increase in the proportion ofmitotic cells in prometaphase/metaphase (Fig. 7C). This suggests thatdespite the decrease in total cells in mitosis, PARG inhibition does effect

Fig. 3. PARG and PARP inhibitors have different effects on repair of ionising radiation-induced DNA damage. (A) Percentage of cells displaying>10 γH2AX foci/cell in untreated(DMSO), PARG inhibited (0.3 μM PDD00017273), or PARP inhibited (1 μM olaparib) MCF-7 cells in the absence and at various times post 3 Gy ionising radiation (IR). Mean and standarderror of the mean of three independent repeats is shown. Statistical significance calculated by two-sided Student’s T-test, cf. peak H2AX activation under same conditions, where *, ** and*** represent p < 0.05, 0.01 and<0.001. (B) Example images of γH2AX foci (Red) co-stained with DAPI. MCF-7 cells were incubated for 16 h with inhibitor before exposure to IR.

P. Gravells et al. DNA Repair 61 (2018) 25–36

30

mitotic progression in IR treated MCF-7 cells. The large shift in theproportion of cells in metaphase was not seen following IR with PARPinhibition, where similar to control, 63% of mitotic cells were in me-taphase. In the absence of IR, PDD00017273 alone did not alter thepercentage of cells in mitosis (Fig. 6B), nor did it effect progressionthrough mitosis (Supplementary Fig. 4), suggesting at the doses usedhere it does not act directly as a spindle poison.

Tankyrases (PARP5a and PARP5b) function during mitosis [75–81].Their action is likely to be reversed by PARG but it is not inhibited byolaparib. The incidence of aberrant mitosis and mitotic progressionwere therefore examined after incubation with a tankyrase inhibitor

[68]. Following IR, the tankyrase inhibitor did phenocopyPDD00017273, in that it increased the incidence of aberrant spindles,however it did not lead to accumulation of cells in metaphase (Sup-plementary Fig. 5). Interestingly inhibition of tankyrases also sensitizedto IR (Supplementary Fig. 6). While this was not to the extent seen withPD00017273, it does suggest that the likely mitotic function of PARG inthe reversal of tankyrase activity may have a role to play in radio-sensitization.

Concurrent with scoring aberrant mitosis, the percentage of cellswith multinucleation was assessed. PARG inhibition also increased 2fold the amount of IR induced multinucleation (Fig. 7E). Approximately

Fig. 4. PARG and PARP inhibitors have different effects on activation of DNA damage repair pathways following ionising radiation. Percentage of cells displaying (A)> 10 DNA-PKcs foci/cell and (C)> 10 RAD51 foci/cell, in untreated (DMSO), PARG inhibited (0.3 μM PDD00017273), or PARP inhibited (1 μM olaparib) MCF-7 cells in the absence and at varioustimes post 3 Gy ionising radiation (IR). Mean and SEM of three independent repeats is shown. Statistical significance calculated by two-sided Student’s T-test, where *, ** and ***represent p < 0.05,< 0.01 and<0.001 respectively In (A) significance is compared with DMSO control at the equivalent time point and in (C) significance calculated compared withthe sample indicated. (B) Example images of DNA-PKcs foci and (D) RAD51 foci (Red) each costained with DAPI. In all cases MCF-7 cells were incubated for 16 h with inhibitor beforeexposure to IR.

P. Gravells et al. DNA Repair 61 (2018) 25–36

31

1000 cells were counted of which no cells were found to be multi-nucleate in the absence of IR regardless of PARP or PARG inhibition(data not shown).

4. Discussion

This is the first report of radiosensitization by a first in class, cellpermeable, specific inhibitor of PARG, PDD00017273, and supports theproposal made by genetic studies [51,62–64] that PARG inhibitorscould be used clinically. In addition, we present the first direct com-parison of the radiosensitizing effects of the PARP1/2/3 inhibitor,olaparib, with a PARG inhibitor. Sensitization to IR was of the same

magnitude and, as expected, both functioned by altering the DNA da-mage response. However, the way in which each sensitized appeared todiffer.

PARP1 and PARP2 function at collapsed replication forks [18,82]and radiosensitization by inhibitors of PARPs is thought to be replica-tion dependent [24,27], with an increase in γH2AX and RAD51 focireported. Similar to these reports, here IR-induced γH2AX persisted atlater times post-IR in olaparib treated compared with control cells and agreater number of RAD51 foci were seen. In addition, there were moreDNA-PKcs foci at later times post-IR in olaparib treated than in controlcells. These data are consistent with the idea that replication associatedDNA breaks are repaired at later times [18] and/or that the presence or

Fig. 5. PARP and PARG inhibitors increase γH2AX positive micro-nuclei after ionising radiation. Micronuclei (MN) frequency in un-treated (DMSO), PARG inhibited (0.3 μM PDD00017273), or PARPinhibited (1 μM olaparib) MCF-7 cells in the absence or 12 h post 3 Gyionising radiation (IR). Mean and SEM of three independent repeats isshown. Statistical significance calculated by two-sided Student’s T-testcompared with DMSO control under equivalent conditions, where *represents p < 0.05. Representative images depicting (i) γH2AX ne-gative and (ii) γH2AX positive MN are shown below. A full data set ofMN under each condition is shown in Supplementary Fig. 3.

Fig. 6. PARP and PARG inhibitors increase G1 arrestand reduce mitotic index following ionising radia-tion. Percentage of cells in each phase of the cellcycle as determined by FACS analysis of PI and pH3stained MCF-7 cells untreated (DMSO), PARG in-hibited (0.3 μM PDD00017273), or PARP inhibited(1 μM olaparib) in the absence or 24 h post 3 Gy io-nising radiation (IR). For clarity (A) depicts allphases of cell cycle and (B) represents only the mi-totic fraction (pH3 positive) from the same data setplotted on a different scale. Mean and SEM of threeindependent repeats is shown. Statistical significancecalculated by two-sided Student’s T-test comparedwith DMSO control under equivalent conditions,where * and *** represent p < 0.05 and<0.001respectively.

P. Gravells et al. DNA Repair 61 (2018) 25–36

32

absence of PARPs 1/2 and 3 can alter the recruitment of the non-homologous end-joining factors XRCC6/XRCC7 (KU70/80) to DSBs[15,83].

In contrast, inhibition of PARG resulted in faster repair of IR-in-duced DNA damage and concomitant rapid activation of significantlyhigher levels of DNA-PKcs. NHEJ involving DNA-PKcs is thought topredominate during G1 phase of the cell cycle, while during S and G2both NHEJ and HR can function [84,85]. Our data suggest that pro-longed activation of PAR can increase recruitment of NHEJ to DNAdamage during G1, and/or alter the balance of NHEJ/HR to DNA da-mage in other phases of the cell cycle. At later time points PARG in-hibitors also led to higher levels of RAD51 foci perhaps indicative of aseparate role for PARG at collapsed replication forks [53,66]. Thisfunctional difference at classical DSBs and replication fork associatedDSBs is supported by that fact that in the absence of any exogenousDNA damage, PARG inhibitors were seen to activate HR but not NHEJ.

There are conflicting reports of the effect of PARG depletion on cellcycle progression, with one report of IR treated HeLa cells having in-creased G2/M arrest and accumulation of cells in metaphase [51],while another in lung (A427) and prostate (PC-14) cancer cell linesdemonstrated suppression of the G2/M checkpoint [64]. Clearly themutational landscape of individual cell lines will affect their response toa PARG inhibitor and/or IR. However, here in MCF-7 breast cancercells, while PARG inhibition resulted in a reduced G2/M population anda reduced mitotic index, those cells in mitosis had an increased in-cidence of aberrant mitotic figures and a higher proportion of mitoticcells were in metaphase than in control cells. Interestingly, PARP

inhibition also reduced the IR-induced G2/M population, however noincrease in aberrant mitotic figures or metaphase was seen, indicatingthat the aberrant mitotic phenotypes were PARG specific. The PARPinhibitor olaparib is considered selective for PARP1/2/3 [67], whilePARG is predicted to reverse the activity of a range of poly(ADP-ri-bosyl)ating enzymes including tankyrases (TNKS/PARP5a and TNKS2/PARP5b). Tankyrases are required for spindle integrity during mitosisthrough PARylation of nuclear mitotic apparatus protein 1 (NUMA1).Thus failure to cleave the PAR from NUMA1 (installed by the tan-kyrases) may give the observed results [86]. Here inhibition of tan-kyrases resulted in an increase in IR-induced aberrant mitotic pheno-types and led to a small increase in radiosensitivity, thus, it is possiblethat the PARG inhibitor specific mitotic phenotypes observed are due tointerruption of tankyrase function. Alternatively, given the role ofPARP3 during mitosis [86], it is possible that preventing addition ofPAR by PARP3 has functional effects that inhibition of PARP3 does not.Finally, aberrance during mitosis could be the result of increased andinappropriate NHEJ.

Using an early moderate pan-PARP inhibitor (5-hydroxyisoquinolin-1-one) and perhaps more relevantly an exogenously expressed trans-dominant PARP-1 DNA binding domain, PARP-1 was demonstrated topositively regulate p53 transactivation function in response to IR andtherefore allow MCF-7 cells to overcome the p53 dependent G1checkpoint [87]. Here, although a predominant IR-induced G1 arrestwas seen in control cells neither olaparib nor PDD00017273 couldovercome this arrest, rather it was augmented. It is possible that dif-ferences in the scheduling of inhibition and IR account for this, such

Fig. 7. PARG inhibited mitotic cells accumulate inmetaphase and feature increased aberrance. (A)Percentage of mitotic MCF-7 cells untreated (DMSO),PARG inhibited (0.3 μM PDD00017273), or PARPinhibited (1 μM olaparib) 24 h post 3 Gy ionisingradiation (IR). Abnormal spindle defects are detectedby immunofluorescent staining for β-tubulin(Green), pericentrin (Red) and DAPI (Blue), thevalue above the bars indicates the percentage of thetotal cell population in mitosis. (B) Representativeimages of aberrant phenotypes seen. (C) Distributionof cells in each phase of mitosis assessed from cellsstained as above. (D) Percentage of multinucleatedcells as stained above, example images shown toright.

P. Gravells et al. DNA Repair 61 (2018) 25–36

33

that the increased pre-incubation time carried out here, allows p53 toovercome PARP/PARG inhibition. Future studies of p53 function inmoderating radiation response upon the PARG/PARG inhibition will beimportant for future clinical application. In MCF-7 cells the number ofcells in G1 is further increased as cells are released from the transientG2/M arrest and pass into G1 [88]. It is tempting to speculate that incells where the DNA damage repair pathways are altered (i.e. followingPARP or PARG inhibition), cells with damaged DNA can persist into G1where they die of mitotic catastrophe. Supportive of this hypothesis wesee increased IR-induced γH2AX positive micronuclei and increasednumbers of multinucleated cells.

5. Conclusion

In summary, previously we demonstrated the use of the PARG in-hibitor PDD00017273 for specific killing of cells defective in certain HRproteins including BRCA1/2 [66]. Here, the same inhibitor is shown toradiosensitize. This is the first report of radiosensitization by a PARGinhibitor and adds to the growing evidence that like PARP, inhibition ofPARG has clinical potential. However, when looking at the mechanismby which sensitization occurs there are clear differences between PARPand PARG inhibition, and it is important that further investigation intothese differences is undertaken.

Author contributions

HB conceived, designed and initiated the study. PG, JN, EG and CDperformed the experiments. AN designed and synthesised the tankyraseinhibitor. DJ and KS designed and synthesised PDD00017273. PG andHB wrote the manuscript. All authors discussed the results and com-mented on the manuscript.

Funding

This work was supported by Breast Cancer Now [grant numberPR016], and CRUK grant [number C5759/A17098]. Microscopy wasperformed on equipment purchased by Welcome Trust grantWT093134AIA and MRC SHIMA award MR/K015753/1.

Conflict of interest

The authors declare no competing financial interest.

Acknowledgements

The authors wish to thank Prof. Mark Meuth, Dr. Ian Waddell andDr. Spencer Collis for their valuable discussions during preparation ofthis work.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, in theonline version, at http://dx.doi.org/10.1016/j.dnarep.2017.11.004.

References

[1] V. Schreiber, F. Dantzer, J.C. Ame, G. de Murcia, Poly(ADP-ribose): novel functionsfor an old molecule, Nat. Rev. Mol. Cell Biol. 7 (2006) 517–528.

[2] R.C. Benjamin, D.M. Gill, Poly(ADP-ribose) synthesis in vitro programmed by da-maged DNA: a comparison of DNA molecules containing different types of strandBreaks, J. Biol. Chem. 255 (1980) 10502–10508.

[3] R. Ding, Y. Pommier, V.H. Kang, M. Smulson, Depletion of poly(ADP-ribose)polymerase by antisense RNA expression results in a delay in DNA strand breakrejoining, J. Biol. Chem. 267 (1992) 12804–12812.

[4] C. Trucco, F.J. Oliver, G. de Murcia, J. Menissier-de Murcia, DNA repair defect inpoly(ADP-ribose) polymerase-deficient cell lines, Nucleic Acids Res. 26 (1998)2644–2649.

[5] F. Le Page, V. Schreiber, C. Dherin, G. De Murcia, S. Boiteux, Poly(ADP-ribose)

polymerase-1(PARP-1) is required in murine cell lines for base excision repair ofoxidative DNA damage in absence of DNA polymerase beta, J. Biol. Chem. 278(2003) 18471–18477.

[6] A.E. Fisher, H. Hochegger, S. Takeda, K.W. Caldecott, Poly(ADP-ribose) polymerase1 accelerates single-strand break repair in concert with poly(ADP-ribose) glycohy-drolase, Mol. Cell. Biol. 27 (2007) 5597–5605.

[7] C.E. Strom, F. Johansson, M. Uhlen, C.A. Szigyarto, K. Erixon, T. Helleday, Poly(ADP-ribose) polymerase (PARP) is not involved in base excision repair but PARPinhibition traps a single-strand intermediate, Nucleic Acids Res. 39 (2011)3166–3175.

[8] S.F. El-Khamisy, M. Masutani, H. Suzuki, K.W. Caldecott, A requirement for PARP-1for the assembly or stability of XRCC1 nuclear foci at sites of oxidative DNA da-mage, Nucleic Acids Res. 31 (2003) 5526–5533.

[9] S. Okano, L. Lan, K.W. Caldecott, T. Mori, A. Yasui, Spatial and temporal cellularresponses to single-strand breaks in human cells, Mol. Cell. Biol. 23 (2003)3974–3981.

[10] S.L. Rulten, F. Cortes-Ledesma, L. Guo, N.J. Iles, K.W. Caldecott, APLF (C2orf13) isa novel component of poly(ADP-ribose) signaling in mammalian cells, Mol. Cell.Biol. 28 (2008) 4620–4628.

[11] C. Morrison, G.C. Smith, L. Stingl, S.P. Jackson, E.F. Wagner, Z.Q. Wang, Geneticinteraction between PARP and DNA-PK in V(D)J recombination and tumorigenesis,Nat. Genet. 17 (1997) 479–482 477.

[12] S. Boulton, S. Kyle, B.W. Durkacz, Interactive effects of inhibitors of poly(ADP-ribose) polymerase and DNA-dependent protein kinase on cellular responses to DNAdamage, Carcinogenesis 20 (1999) 199–203.

[13] M. Audebert, B. Salles, P. Calsou, Involvement of poly(ADP-ribose) polymerase-1and XRCC1/DNA ligase III in an alternative route for DNA double-strand breaksrejoining, J. Biol. Chem. 279 (2004) 55117–55126.

[14] S.L. Rulten, A.E. Fisher, I. Robert, M.C. Zuma, M. Rouleau, L. Ju, G. Poirier,B. Reina-San-Martin, K.W. Caldecott, PARP-3 and APLF function together to ac-celerate nonhomologous end-joining, Mol. Cell 41 (2011) 33–45.

[15] C. Beck, C. Boehler, J. Guirouilh Barbat, M.E. Bonnet, G. Illuzzi, P. Ronde,L.R. Gauthier, N. Magroun, A. Rajendran, B.S. Lopez, et al., PARP3 affects the re-lative contribution of homologous recombination and nonhomologous end-joiningpathways, Nucleic Acids Res. 42 (2014) 5616–5632.

[16] Y.G. Yang, U. Cortes, S. Patnaik, M. Jasin, Z.Q. Wang, Ablation of PARP-1 does notinterfere with the repair of DNA double-strand breaks, but compromises the re-activation of stalled replication forks, Oncogene 23 (2004) 3872–3882.

[17] K. Sugimura, S. Takebayashi, H. Taguchi, S. Takeda, K. Okumura, PARP-1 ensuresregulation of replication fork progression by homologous recombination on da-maged DNA, J. Cell Biol. 183 (2008) 1203–1212.

[18] H.E. Bryant, E. Petermann, N. Schultz, A.S. Jemth, O. Loseva, N. Issaeva,F. Johansson, S. Fernandez, P. McGlynn, T. Helleday, PARP is activated at stalledforks to mediate Mre11-dependent replication restart and recombination, EMBO J.28 (2009) 2601–2615.

[19] H.E. Bryant, N. Schultz, H.D. Thomas, K.M. Parker, D. Flower, E. Lopez, S. Kyle,M. Meuth, N.J. Curtin, T. Helleday, Specific killing of BRCA2-deficient tumourswith inhibitors of poly(ADP-ribose) polymerase, Nature 434 (2005) 913–917.

[20] A. Drean, C.J. Lord, A. Ashworth, PARP inhibitor combination therapy, Crit. Rev.Oncol. Hematol. 108 (2016) 73–85.

[21] S. Shall, G. de Murcia, Poly(ADP-ribose) polymerase-1: what have we learned fromthe deficient mouse model? Mutat. Res. 460 (2000) 1–15.

[22] J.M. deMurcia, C. Niedergang, C. Trucco, M. Ricoul, B. Dutrillaux, M. Mark,F.J. Oliver, M. Masson, A. Dierich, M. LeMeur, et al., Requirement of poly(ADP-ribose) polymerase in recovery from DNA damage in mice and in cells, Proc. Nat.Acad. Sci. U. S. A. 94 (1997) 7303–7307.

[23] E. Benhur, C.C. Chen, M.M. Elkind, Inhibitors of Poly(Adenosine diphosphoribose)synthetase, examination of metabolic perturbations, and enhancement of radiationresponse in chinese-hamster cells, Cancer Res. 45 (1985) 2123–2127.

[24] C. Godon, F.P. Cordelieres, D. Biard, N. Giocanti, F. Megnin-Chanet, J. Hall,V. Favaudon, PARP inhibition versus PARP-1 silencing: different outcomes in termsof single-strand break repair and radiation susceptibility, Nucleic Acids Res. 36(2008) 4454–4464.

[25] A. Chalmers, P. Johnston, M. Woodcock, M. Joiner, B. Marples, PARP-1, PARP-2,and the cellular response to low doses of ionizing radiation, Int. J. Radiat. Oncol. 58(2004) 410–419.

[26] A. Schlicker, P. Peschke, A. Burkle, E.W. Hahn, J.H. Kim, 4-amino-1,8-naphthali-mide: a novel inhibitor of poly(ADP-ribose) polymerase and radiation sensitizer,Int. J. Radiat. Biol. 75 (1999) 91–100.

[27] F.A. Dungey, D.A. Loser, A.J. Chalmers, Replication-dependent radiosensitization ofhuman glioma cells by inhibition of poly(Adp-Ribose) polymerase: mechanisms andtherapeutic potential, Int. J. Radiat. Oncol. 72 (2008) 1188–1197.

[28] K.A. Bridges, C. Toniatti, C.A. Buser, H. Liu, T.A. Buchholz, R.E. Meyn, Niraparib(MK-4827), a novel poly(ADP-Ribose) polymerase inhibitor, radiosensitizes humanlung and breast cancer cells, Oncotarget 5 (2014) 5076–5086.

[29] E.V. Efimova, H.J. Mauceri, D.W. Golden, E. Labay, V.P. Bindokas, T.E. Darga,C. Chakraborty, J.C. Barreto-Andrade, C. Crawley, H.G. Sutton, et al., Poly(ADP-ribose) polymerase inhibitor induces accelerated senescence in irradiated breastcancer cells and tumors, Cancer Res. 70 (2010) 6277–6282.

[30] F.Y. Feng, C. Speers, M. Liu, W.C. Jackson, D. Moon, J. Rinkinen, K. Wilder-Romans,R. Jagsi, L.J. Pierce, Targeted radiosensitization with PARP1 inhibition: optimiza-tion of therapy and identification of biomarkers of response in breast cancer, BreastCancer Res. Treat. 147 (2014) 81–94.

[31] L. Wang, K.A. Mason, K.K. Ang, T. Buchholz, D. Valdecanas, A. Mathur, C. Buser-Doepner, C. Toniatti, L. Milas, MK-4827, a PARP-1/-2 inhibitor, strongly enhancesresponse of human lung and breast cancer xenografts to radiation, Invest. New

P. Gravells et al. DNA Repair 61 (2018) 25–36

34

Drugs 30 (2012) 2113–2120.[32] L. Wang, K.A. Mason, K.K. Ang, T. Buchholz, D. Valdecanas, A. Mathur, C. Buser-

Doepner, C. Toniatti, L. Milas, MK-4827, a PARP-1/-2 inhibitor, strongly enhancesresponse of human lung and breast cancer xenografts to radiation, Invest. New Drug30 (2012) 2113–2120.

[33] M. Ali, M. Kamjoo, H.D. Thomas, S. Kyle, I. Pavlovska, M. Babur, B.A. Telfer,N.J. Curtin, K.J. Williams, The clinically active PARP inhibitor AG014699 ame-liorates cardiotoxicity but does not enhance the efficacy of doxorubicin, despiteimproving tumor perfusion and radiation response in mice, Mol. Cancer Ther. 10(2011) 2320–2329.

[34] C.K. Donawho, Y. Luo, Y.P. Luo, T.D. Penning, J.L. Bauch, J.J. Bouska,V.D. Bontcheva-Diaz, B.F. Cox, T.L. DeWeese, L.E. Dillehay, et al., ABT-888, anorallyactive poly(ADP-ribose) polymerase inhibitor that potentiates DNA-damagingagents in preclinical tumor models, Clin. Cancer Res. 13 (2007) 2728–2737.

[35] G.D. Zhu, J.C. Gong, V.B. Gandhi, Y. Luo, X.S. Liu, Y. Shi, V. Klinghofer,E.F. Johnson, D. Frost, C. Donawho, et al., Discovery and SAR of ABT-888, an in-hibitor of poly(ADP-ribose) polymerase (PARP) for the treatment of cancer, Abstr.Pap. Am. Chem. Soc. 233 (2007) 716-716.

[36] C.R. Calabrese, R. Almassy, S. Barton, M.A. Batey, A.H. Calvert, S. Canan-Koch,B.W. Durkacz, Z. Hostomsky, R.A. Kumpf, S. Kyle, et al., Anticancer chemosensi-tization and radiosensitization by the novel poly(ADP-ribose) polymerase-1 in-hibitor AG14361, J. Nat. Cancer. Inst. 96 (2004) 56–67.

[37] J.W. Shelton, T.V. Waxweiler, J. Landry, H.Y. Gao, Y.B. Xu, L.F. Wang, B. El-Rayes,H.K.G. Shu, In vitro and In vivo enhancement of chemoradiation using the oralPARP inhibitor ABT-888 in colorectal cancer cells, Int. J. Radiat. Oncol. 86 (2013)469–476.

[38] J.M. Albert, C. Cao, K.W. Kim, C.D. Willey, L. Geng, D.K. Xiao, H. Wang, A. Sandler,D.H. Johnson, A.D. Colevas, et al., Inhibition of poly (ADP-ribose) polymerase en-hances cell death and improves tumor growth delay in irradiated lung cancermodels, Clin. Cancer Res. 13 (2007) 3033–3042.

[39] J.M. Senra, B.A. Telfer, K.E. Cherry, C.M. McCrudden, D.G. Hirst, M.J. O’Connor,S.R. Wedge, I.J. Stratford, Inhibition of PARP-1 by olaparib (AZD2281) increasesthe radiosensitivity of a lung tumor xenograft, Mol. Cancer Ther. 10 (2011)1949–1958.

[40] J.P.H. Chow, W.Y. Man, M. Mao, H. Chen, F. Cheung, J. Nicholls, S.W. Tsao,M.L. Lung, R.Y.C. Poon, PARP1 is overexpressed in nasopharyngeal carcinoma andits inhibition enhances radiotherapy, Mol. Cancer Ther. 12 (2013) 2517–2528.

[41] K. Khan, K. Araki, D.Y. Wang, G.Y. Li, X. Li, J. Zhang, W.Z. Xu, R.K. Hoover,S. Lauter, B. O'Malley, et al., Head and neck cancer radiosensitization by the novelpoly(Adp-Ribose) polymerase inhibitor gpi-15427, Head Neck-J. Sci. Specialities 32(2010) 381–391.

[42] M. Venere, P. Hamerlik, Q. Wu, R.D. Rasmussen, L.A. Song, A. Vasanji, N. Tenley,W.A. Flavahan, A.B. Hjelmeland, J. Bartek, et al., Therapeutic targeting of con-stitutive PARP activation compromises stem cell phenotype and survival of glio-blastoma-initiating cells, Cell Death Differ. 21 (2014) 258–269.

[43] S. Mueller, S. Bhargava, A.M. Molinaro, X.D. Yang, I. Kolkowitz, A. Olow,N. Wehmeijer, S. Orbach, J. Chen, K.K. Matthay, et al., Poly (ADP-Ribose) poly-merase inhibitor MK-4827 together with radiation as a novel therapy for metastaticneuroblastoma, Anticancer Res. 33 (2013) 755–762.

[44] G. Noel, C. Godon, M. Fernet, N. Giocanti, F. Megnin-Chanet, V. Favaudon,Radiosensitization by the poly(ADP = ribose) polymerase inhibitor 4-amino-1,8-naphthalimide is specific of the S phase of the cell cycle and involves arrest of DNAsyntheses, Mol. Cancer Ther. 5 (2006) 564–574.

[45] F.A. Dungey, D.A. Loser, A.J. Chalmers, Replication-dependent radiosensitization ofhuman glioma cells by inhibition of poly(ADP-Ribose) polymerase: mechanisms andtherapeutic potential, Int. J. Radiat. Oncol. Biol. Phys. 72 (2008) 1188–1197.

[46] M. Miwa, M. Tanaka, T. Matsushima, T. Sugimura, Purification and properties ofglycohydrolase from calf thymus splitting ribose–ribose linkages of poly(adenosinediphosphate ribose), J. Biol. Chem. 249 (1974) 3475–3482.

[47] S.A. Braun, P.L. Panzeter, M.A. Collinge, F.R. Althaus, Endoglycosidic cleavage ofbranched polymers by poly(ADP-ribose) glycohydrolase, Eur. J. Biochem. 220(1994) 369–375.

[48] G. Brochu, C. Duchaine, L. Thibeault, J. Lagueux, G.M. Shah, G.G. Poirier, Mode ofaction of poly(ADP-ribose) glycohydrolase, Biochim. Biophys. Acta 1219 (1994)342–350.

[49] D. Slade, M.S. Dunstan, E. Barkauskaite, R. Weston, P. Lafite, N. Dixon, M. Ahel,D. Leys, I. Ahel, The structure and catalytic mechanism of a poly(ADP-ribose)glycohydrolase, Nature 477 (2016) 616–620.

[50] E. Barkauskaite, G. Jankevicius, I. Ahel, Structures and mechanisms of enzymesemployed in the synthesis and degradation of PARP-Dependent protein ADP-ribo-sylation, Mol. Cell 58 (2015) 935–946.

[51] J.C. Ame, E. Fouquerel, L.R. Gauthier, D. Biard, F.D. Boussin, F. Dantzer, G. deMurcia, V. Schreiber, Radiation-induced mitotic catastrophe in PARG-deficientcells, J. Cell Sci. 122 (2009) 1990–2002.

[52] K. Erdelyi, P. Bai, I. Kovacs, E. Szabo, G. Mocsar, A. Kakuk, C. Szabo, P. Gergely,L. Virag, Dual role of poly(ADP-ribose) glycohydrolase in the regulation of celldeath in oxidatively stressed A549 cells, FASEB J. 23 (2009) 3553–3563.

[53] C. Fathers, R.M. Drayton, S. Solovieva, H.E. Bryant, Inhibition of poly(ADP-ribose)glycohydrolase (PARG) specifically kills BRCA2-deficient tumor cells, ABBV CellCycle 11 (2012) 990–997.

[54] A. Ray Chaudhuri, A.K. Ahuja, R. Herrador, M. Lopes, Poly(ADP-ribosyl) glycohy-drolase prevents the accumulation of unusual replication structures during un-perturbed S phase, Mol. Cell. Biol. 35 (2015) 856–865.

[55] G. Illuzzi, E. Fouquerel, J.C. Ame, A. Noll, K. Rehmet, H.P. Nasheuer, F. Dantzer,V. Schreiber, PARG is dispensable for recovery from transient replicative stress butrequired to prevent detrimental accumulation of poly(ADP-ribose) upon prolonged

replicative stress, Nucleic Acids Res. 42 (2014) 7776–7792.[56] W. Min, U. Cortes, Z. Herceg, W.M. Tong, Z.Q. Wang, Deletion of the nuclear iso-

form of poly(ADP-ribose) glycohydrolase (PARG) reveals its function in DNA repair,genomic stability and tumorigenesis, Carcinogenesis 31 (2010) 2058–2065.

[57] X. Feng, D.W. Koh, Inhibition of poly(ADP-ribose) polymerase-1 or poly(ADPribose)glycohydrolase individually, but not in combination, leads to improved che-motherapeutic efficacy in HeLa cells, Int. J. Oncol. 42 (2013) 749–756.

[58] H. Shirai, A.R. Poetsch, A. Gunji, D. Maeda, H. Fujimori, H. Fujihara, T. Yoshida,H. Ogino, M. Masutani, PARG dysfunction enhances DNA double strand breakformation in S-phase after alkylation DNA damage and augments different celldeath pathways, Cell. Death Dis. 4 (2013) e656.

[59] X. Feng, Y. Zhou, A.M. Proctor, M.M. Hopkins, M. Liu, D.W. Koh, Silencing ofApoptosis-Inducing factor and poly(ADP-ribose) glycohydrolase reveals novel rolesin breast cancer cell death after chemotherapy, Mol. Cancer 11 (2012) 48–56.

[60] C. Blenn, F.R. Althaus, M. Malanga, Poly(ADP-ribose) glycohydrolase silencingprotects against H2O2-induced cell death, Biochem. J. 396 (2006) 419–429.

[61] H.-Y. Huang, J.-F. Cai, Q.-C. Liu, G.-H. Hu, B. Xia, J.-Y. Mao, D.-S. Wu, J.-J. Liu, Z.-X. Zhuang, Role of poly(ADP-ribose) glycohydrolase in the regulation of cell fate inresponse to benzo(a)pyrene, Exp. Cell Res. 318 (2012) 682–690.

[62] U. Cortes, W.M. Tong, D.L. Coyle, M.L. Meyer-Ficca, R.G. Meyer, V. Petrilli,Z. Herceg, E.L. Jacobson, M.K. Jacobson, Z.Q. Wang, Depletion of the 110-kilo-dalton isoform of poly(ADP-ribose) glycohydrolase increases sensitivity to geno-toxic and endotoxic stress in mice, Mol. Cell. Biol. 24 (2004) 7163–7178.

[63] H. Shirai, H. Fujimori, A. Gunji, D. Maeda, T. Hirai, A.R. Poetsch, H. Harada,T. Yoshida, K. Sasai, R. Okayasu, et al., Parg deficiency confers radio-sensitizationthrough enhanced cell death in mouse ES cells exposed to various forms of ionizingradiation, Biochem. Biophys. Res. Commun. 435 (2013) 100–106.

[64] Y. Nakadate, Y. Kodera, Y. Kitamura, T. Tachibana, T. Tamura, F. Koizumi,Silencing of poly(ADP-ribose) glycohydrolase sensitizes lung cancer cells to radia-tion through the abrogation of DNA damage checkpoint, Biochem. Biophys. Res.Commun. 441 (2013) 793–798.

[65] D.I. James, K.M. Smith, A.M. Jordan, E.E. Fairweather, L.A. Griffiths,N.S. Hamilton, J.R. Hitchin, C.P. Hutton, S. Jones, P. Kelly, et al., First-in-classchemical probes against poly(ADP-ribose) glycohydrolase (PARG) inhibit DNA re-pair with differential pharmacology to olaparib, ACS Chem. Biol. 11 (2016)3179–3190.

[66] P. Gravells, E. Grant, K.M. Smith, D.I. James, H.E. Bryant, Specific killing of DNAdamage-response deficient cells with inhibitors of poly(ADP-ribose) glycohydrolase,DNA Repair 52 (2017) 81–91.

[67] L. Oplustil O'Connor, S.L. Rulten, A.N. Cranston, R. Odedra, H. Brown, J.E. Jaspers,L. Jones, C. Knights, B. Evers, A. Ting, et al., The PARP inhibitor AZD2461 providesinsights into the role of PARP3 inhibition for both synthetic lethality and toler-ability with chemotherapy in preclinical models, Cancer Res. 76 (2016) 6084–6094.

[68] A. Nathubhai, T. Haikarainen, J. Koivunen, S. Murthy, F. Koumanov, M.D. Lloyd,G.D. Holman, T. Pihlajaniemi, D. Tosh, L. Lehtio, et al., Highly potent and isoformselective dual site binding Tankyrase/Wnt signaling inhibitors that increase cellularglucose uptake and have antiproliferative activity, J. Med. Chem. 60 (2017)814–820.

[69] M.E. Lomax, L.K. Folkes, P. O'Neill, Biological consequences of radiation-inducedDNA damage: relevance to radiotherapy, Clin. Oncol. 25 (2013) 578–585.

[70] H.E. Bryant, N. Schultz, H.D. Thomas, K.M. Parker, D. Flower, E. Lopez, S. Kyle,M. Meuth, N.J. Curtin, T. Helleday, Specific killing of BRCA2-deficient tumourswith inhibitors of poly(ADP-ribose) polymerase, Nature 434 (2005) 913–917.

[71] J.S. Kim, T.B. Krasieva, H. Kurumizaka, D.J. Chen, A.M. Taylor, K. Yokomori,Independent and sequential recruitment of NHEJ and HR factors to DNA damagesites in mammalian cells, J. Cell Biol. 170 (2005) 341–347.

[72] N.G. Medvedeva, I.V. Panyutin, I.G. Panyutin, R.D. Neumann, Phosphorylation ofhistone H2AX in radiation-induced micronuclei, Radiat. Res. 168 (2007) 493–498.

[73] T. Yoshikawa, G. Kashino, K. Ono, M. Watanabe, Phosphorylated H2AX foci intumor cells have No correlation with their radiation sensitivities, J. Radiat. Res. 50(2009) 151–160.

[74] M. Fabbro, K. Savage, K. Hobson, A.J. Deans, S.N. Powell, G.A. McArthur,K.K. Khanna, BRCA1-BARD1 complexes are required for p53Ser-15 phosphorylationand a G1/S arrest following ionizing radiation-induced DNA damage, J. Biol. Chem.279 (2004) 31251–31258.

[75] P. Chang, M. Coughlin, T.J. Mitchison, Tankyrase-1 polymerization of poly(ADP-ribose) is required for spindle structure and function, Nat. Cell Biol. 7 (2005)1133–1139.

[76] P. Chang, M. Coughlin, T.J. Mitchison, Interaction between Poly(ADP-ribose) andNuMA contributes to mitotic spindle pole assembly, Mol. Biol. Cell 20 (2009)4575–4585.

[77] W. Chang, J.N. Dynek, S. Smith, NuMA is a major acceptor of poly(ADP-ribosyl)ation by tankyrase 1 in mitosis, Biochem. J 391 (2005) 177–184.

[78] G.H. Ha, H.S. Kim, H. Go, H. Lee, H. Seimiya, D.H. Chung, C.W. Lee, Tankyrase-1function at telomeres and during mitosis is regulated by Polo-like kinase-1-medi-ated phosphorylation, Cell Death Differ. 19 (2012) 321–332.

[79] Y. Ozaki, H. Matsui, H. Asou, A. Nagamachi, D. Aki, H. Honda, S. Yasunaga,Y. Takihara, T. Yamamoto, S. Izumi, et al., Poly-ADP ribosylation of Miki by tan-kyrase-1 promotes centrosome maturation, Mol. Cell 47 (2012) 694–706.

[80] E. Tripathi, S. Smith, Cell cycle-regulated ubiquitination of tankyrase 1 by RNF8and ABRO1/BRCC36 controls the timing of sister telomere resolution, EMBO J. 36(2017) 503–519.

[81] T.Y. Yeh, J.I. Sbodio, N.W. Chi, Mitotic phosphorylation of tankyrase, a PARP thatpromotes spindle assembly, by GSK3, Biochem. Biophys. Res. Commun. 350 (2006)574–579.

[82] N. Saleh-Gohari, H.E. Bryant, N. Schultz, K.M. Parker, T.N. Cassel, T. Helleday,

P. Gravells et al. DNA Repair 61 (2018) 25–36

35

Spontaneous homologous recombination is induced by collapsed replication forksthat are caused by endogenous DNA single-strand breaks, Mol. Cell. Biol. 25 (2005)7158–7169.

[83] M.L. Wang, W.Z. Wu, W.Q. Wu, B. Rosidi, L.H. Zhang, H.C. Wang, G. Iliakis, PARP-1and Ku compete for repair of DNA double strand breaks by distinct NHEJ pathways,Nucleic Acids Res. 34 (2006) 6170–6182.

[84] K. Rothkamm, I. Kruger, L.H. Thompson, M. Lobrich, Pathways of DNA double-strand break repair during the mammalian cell cycle, Mol. Cell. Biol. 23 (2003)5706–5715.

[85] S.E. Lee, R.A. Mitchell, A. Cheng, E.A. Hendrickson, Evidence for DNA-PK-depen-dent and – independent DNA double-strand break repair pathways in mammalian

cells as a function of the cell cycle, Mol. Cell. Biol. 17 (1997) 1425–1433.[86] C. Boehler, L.R. Gauthier, O. Mortusewicz, D.S. Biard, J.M. Saliou, A. Bresson,

S. Sanglier-Cianferani, S. Smith, V. Schreiber, F. Boussin, et al., Poly(ADP-ribose)polymerase 3 (PARP3), a newcomer in cellular response to DNA damage and mi-totic progression, Proc. Nat. Acad. Sci. U. S. A. 108 (2011) 2783–2788.

[87] S. Wieler, J.P. Gagne, H. Vaziri, G.G. Poirier, Poly(ADP-ribose) polymerase-1 is apositive regulator of the p53-mediated G(1) arrest response following ionizing ra-diation, J. Biol. Chem. 278 (2003) 18914–18921.

[88] J.M.C. Martin, A. Balkenende, T. Verschoor, F. Lallemand, R. Michalides, Cyclin D1overexpression enhances radiation-induced apoptosis and radiosensitivity in abreast tumor cell line, Cancer Res. 59 (1999) 1134–1140.

P. Gravells et al. DNA Repair 61 (2018) 25–36

36